A dual-task segmentation network based on multi-head hierarchical attention for 3D plant point cloud

Na minha lista:
Detalhes bibliográficos
Publicado no:Frontiers in Plant Science vol. 16 (Jul 2025), p. 1610443-1610459
Autor principal: Pan, Dan
Outros Autores: Liu, Baijing, Luo, Lin, Zeng, An, Zhou, Yuting, Pan, Kaixin, Xian, Zhiheng, Yulun Xian, Liu, Licheng
Publicado em:
Frontiers Media SA
Assuntos:
Acesso em linha:Citation/Abstract
Full Text
Full Text - PDF
Tags: Adicionar Tag
Sem tags, seja o primeiro a adicionar uma tag!

MARC

LEADER 00000nab a2200000uu 4500
001 3273795140
003 UK-CbPIL
022 |a 1664-462X 
024 7 |a 10.3389/fpls.2025.1610443  |2 doi 
035 |a 3273795140 
045 2 |b d20250701  |b d20250731 
100 1 |a Pan, Dan  |u School of Electronics and Information, Guangdong Polytechnic Normal University, Guangzhou, China 
245 1 |a A dual-task segmentation network based on multi-head hierarchical attention for 3D plant point cloud 
260 |b Frontiers Media SA  |c Jul 2025 
513 |a Journal Article 
520 3 |a IntroductionThe development of automated high-throughput plant phenotyping systems with non-destructive characteristics fundamentally relies on achieving accurate segmentation of botanical structures at both semantic and instance levels. However, most existing approaches rely heavily on empirically determined threshold parameters and rarely integrate semantic and instance segmentation within a unified framework.MethodsTo address these limitations, this study introduces a methodology leveraging 2D image data of real plants, i.e., Caladium bicolor, captured using a custom-designed plant cultivation platform. A high-quality 3D point cloud dataset was generated through reconstruction. Building on this foundation, we propose a streamlined Dual-Task Segmentation Network (DSN) incorporating a multi-head hierarchical attention mechanism to achieve superior segmentation performance. Also, the dual-task framework employs Multi-Value Conditional Random Field (MV-CRF) to enable semantic segmentation of stem-leaf and individual leaf identification through the DSN architecture when processing manually-annotated 3D point cloud data. The network features a dual-branch architecture: one branch predicts the semantic class of each point, while the other embeds points into a high-dimensional vector space for instance clustering. Multi-task joint optimization is facilitated through the MV-CRF model.Results and discussionBenchmark evaluations validate the novel framework’s segmentation efficacy, yielding 99.16% macro-averaged precision, 95.73% class-wise recognition rate, and an average Intersection over Union of 93.64%, while comparative analyses confirm its superiority over nine benchmark architectures in 3D point cloud analytics. For instance segmentation, the model achieved leading metrics of 87.94%, 72.36%, and 71.61%, respectively. Furthermore, ablation studies validated the effectiveness of the network’s design and substantiated the rationale behind each architectural choice. 
653 |a Accuracy 
653 |a Comparative analysis 
653 |a Datasets 
653 |a Deep learning 
653 |a Conditional random fields 
653 |a Ablation 
653 |a Leaves 
653 |a Semantic segmentation 
653 |a Registration 
653 |a Automation 
653 |a Plants (botany) 
653 |a Genotype & phenotype 
653 |a Benchmarks 
653 |a Efficiency 
653 |a Vector spaces 
653 |a Image reconstruction 
653 |a Image segmentation 
653 |a Computer vision 
653 |a Phenotyping 
653 |a Clustering 
653 |a Three dimensional models 
653 |a Effectiveness 
653 |a Instance segmentation 
653 |a Plant growth 
653 |a Annotations 
653 |a Morphology 
653 |a Semantics 
653 |a Environmental 
700 1 |a Liu, Baijing  |u School of Information Engineering, Guangdong University of Technology, Guangzhou, China 
700 1 |a Luo, Lin  |u School of Computer Science and Technology, Guangdong University of Technology, Guangzhou, China 
700 1 |a Zeng, An  |u School of Computer Science and Technology, Guangdong University of Technology, Guangzhou, China 
700 1 |a Zhou, Yuting  |u School of Electronics and Information, Guangdong Polytechnic Normal University, Guangzhou, China 
700 1 |a Pan, Kaixin  |u School of Electronics and Information, Guangdong Polytechnic Normal University, Guangzhou, China 
700 1 |a Xian, Zhiheng  |u Guangzhou Huitong Agricultural Technology Co., Ltd., Guangzhou, China 
700 1 |a Yulun Xian  |u Guangzhou Huitong Agricultural Technology Co., Ltd., Guangzhou, China, Guangzhou iGrowLite Agricultural Technology Co., Ltd., Guangzhou, China 
700 1 |a Liu, Licheng  |u School of Information Engineering, Guangdong University of Technology, Guangzhou, China 
773 0 |t Frontiers in Plant Science  |g vol. 16 (Jul 2025), p. 1610443-1610459 
786 0 |d ProQuest  |t Agriculture Science Database 
856 4 1 |3 Citation/Abstract  |u https://www.proquest.com/docview/3273795140/abstract/embedded/H09TXR3UUZB2ISDL?source=fedsrch 
856 4 0 |3 Full Text  |u https://www.proquest.com/docview/3273795140/fulltext/embedded/H09TXR3UUZB2ISDL?source=fedsrch 
856 4 0 |3 Full Text - PDF  |u https://www.proquest.com/docview/3273795140/fulltextPDF/embedded/H09TXR3UUZB2ISDL?source=fedsrch